Seasonal weather forecasts are a frontier product, for which actual operational uses by electric utilities nationwide are not yet known. The objective of this activity is to study seasonal forecasts and their potential use on the Italian territory, interpreting them in an electro-energy context to assess their contribution to the management, operation and planning activities of the electricity system.

In fact, it is possible to think about applying such forecasts to activities that are of growing interest to electricity operators. For example, the identification well in advance of adverse environmental conditions that may lead to critical issues in the management of the electricity system, so that the most vulnerable areas can be identified and appropriate risk mitigation measures can be prepared.

This report describes the latest in terms of seasonal forecasts, based mainly on what has been documented by the European Centre for Medium-Range Weather Forecasts (ECMWF). In general, weather forecasts are created using numerical models to calculate the evolution of the state of the atmosphere, oceans, and processes concerning the Earth’s surface from an initial state based on observations. However, due to the limitations of the observing system, the initial state is not perfectly known.

Since the atmosphere is a chaotic system, the evolution of a forecast is very sensitive to small errors in the initial conditions, and this limits the ability to forecast daily weather variations beyond 10 to 15 days ahead. However, it is possible to provide longer-term forecasts for weeks, months (monthly and seasonal forecasts) and years (climate models) ahead by considering a number of components of the Earth system that evolve more slowly than the atmospheric system, such as the oceans and cryosphere. In fact, due to their slower evolution, they retain information longer, from their initial state, and it is possible to forecast their evolution over longer time scales.

The seasonal forecast is basically a statistical summary of the daily weather calculated from the forecast model for the months ahead. For such extended time horizons, it is in fact not possible to expect the provision of deterministic values relating to an expected quantity for a certain meteorological variable: the most widely used systems consist of an ensemble of different atmospheric evolutions, from which it is possible to derive mean values and calculate expected anomalies for the seasons to come on the basis of historical periods for which a reference model climatology is available.